Tennis: The US Open are the mens/womens draws not so random

[indypacers]

I watched ESPN's "outside the lines" today and they touched on the random number generatr used by the USTA in seeding for the US Open, seems that the first and second seeds had some very runs to a championship due in part to some favorable selection.

I've often wondered that about the USFA draws for seedings, pools and so on. It apears that some fencers just have a way easier time(path) than some fencers. Just compare the strength of pool at the next NAC and give me your thoughts.

[jeremyb215]

I watched ESPN's "outside the lines" today and they touched on the random number generatr used by the USTA in seeding for the US Open, seems that the first and second seeds had some very runs to a championship due in part to some favorable selection.

I've often wondered that about the USFA draws for seedings, pools and so on. It apears that some fencers just have a way easier time(path) than some fencers. Just compare the strength of pool at the next NAC and give me your thoughts.

From what I understand, generating truely random number sequences is a lot more diffacult than it sounds.

[eac]

It doesn't need to be truly random; it just needs to be arbitrary and the next number needs to be uncorrelated with the previous history of numbers.

[Gav]

That's because seeding isn't random.

Have you been to a fencing tournament?

It's not random in Tennis either. The key is in the word "seeding".

[eac]

That's because seeding isn't random.

Have you been to a fencing tournament?

It's not random in Tennis either. The key is in the word "seeding".

I believe that despite his inarticulation he was referring to the randomness required to break large groups of ties in pre-pool seeding, e.g. between the millions of A11's in a D1ME event. The randomness there does have a big effect on results.

But, as has been established on here many times, the rating system does not provide a highly granular or highly current estimate of relative strength, and so randomness for pools is required. The recent addition of sub-points-list-points for 33-64 is supposed to help with that somewhat.

What say we don't do the ratings thread thing, though, search for it, and remember that we did already. IIRC one was called 'Ratings thread.'

[tbryan]

I watched ESPN's "outside the lines" today and they touched on the random number generatr used by the USTA in seeding for the US Open, seems that the first and second seeds had some very runs to a championship due in part to some favorable selection.

I wasn't quite sure what you were talking about. A link to the article may help others: http://espn.go.com/espn/otl/story/_/...tically-likely

"the top two men's and women's seeds in the U.S. Open -- on average -- faced easier opponents in the first round than is statistically probable if the draws were truly random....A truly random draw for the unseeded players -- which is promised by USTA officials..."

I didn't look much at the analysis, but the issue is different from the US Fencing tournament/pool seeding issue. In our case, we have large groups of fencers with an equivalent rating. It seems reasonable to assume that some of those equivalently-rated fencers are actually much better than others. Even with a perfectly random distribution of those fencers across the pools, some pools will be more difficult than others.

In the USTA case, wildcards are apparently supposed to be seeded randomly (not based on their rank). The article claims that the #1 and #2 seed at the US Open end up with easier opponents across multiple years than they should if the wildcard seeding were truly random. Their analysis did not find the same problem when looking at the Australia Open.

Other than "some paths through the DE table are easier," I don't see what the USTA issue has to do with seeding at fencing NACs.

[edew]

I believe that despite his inarticulation he was referring to the randomness required to break large groups of ties in pre-pool seeding, e.g. between the millions of A11's in a D1ME event. The randomness there does have a big effect on results.

But, as has been established on here many times, the rating system does not provide a highly granular or highly current estimate of relative strength, and so randomness for pools is required. The recent addition of sub-points-list-points for 33-64 is supposed to help with that somewhat.

What say we don't do the ratings thread thing, though, search for it, and remember that we did already. IIRC one was called 'Ratings thread.'

The fencers are also seeded based on their FIE and National points before the classification. But I think the randomness issue comes in after the pools when you have 12 people with the same V/B Ind and TS. Then, you could be 3rd or your could be 15th. I think those should also use the FIE and National points as a bias to complete randomness. In other words, if you were highly ranked and you're tied with several others for 3rd coming out of the pools, then you should be seeded higher, in a quasi-random sort, than the lower ranks.

[prototoast]

While there may be some software quirks that make the seeding not truly random, I certainly don't believe that anyone is deliberately manipulating them to affect the outcome. With the quick turnaround, there simply isn't enough time for tournament officials to not only manipulate that which is supposed to be random, but also hide their work from the 50 fencers peering over their shoulders to try to get the first glimpse of the seed.

[erooMynohtnA]

Momma always said "random is as random does."